By Ashok Vemuri, Member of the Executive Council (EC), Head of Americas and Global Head, Financial Services, Infosys Limited


The other day, I was reading an article in a technology journal that offered a glimpse into the cache of data held by various departments within the US Government.


Surprisingly, there is no limitation on the release of this data—whether it be 401(k) performance of US corporations or statistics released by HUD—the United States government is sitting on a booty of data ripe to be analyzed and mobilized for the benefit of society.


What’s preventing the release of this data? Well, it’s not that someone or something is preventing the release of this data; instead, the ability to share data is hindered by the organization of the various government entities, as well as the lack of a policy driving the centralization of data. In short, data is stuck in siloes. Sound familiar?


Reading this piece provided an opportunity to reflect on financial institutions of today. In today’s bank, how much information is shared across channels? What is the CIO doing today to ensure customer data is thoroughly vetted and, to turn a noun into a verb, “actionized?” Is a Unified View of the Customer merely a pipedream? What data is actually captured? In this article, we will show the value of sharing data across the enterprise as well as the 3 Ps of data management—prepare, process, and present.



Sharing was probably the first interpersonal skill you learned as a child. Sharing enabled you to navigate the playroom environment, play fair with other children, and distribute the benefit of a toy to all of your peers. Just so, in a bank, sharing data is as important if not more important than sharing is for a toddler. While the inability of a bank’s lines of businesses to share will not lead to the inevitable tear-filled breakdown, it will, instead, prevent a line of business from fully understanding the needs of the customer it is trying to serve—especially if critical data is trapped in the silo of a disparate line of business. Ultimately, the lack of a unified customer view will result in a below-average customer experience.


In addition to a poor customer experience, unshared, siloed data decreases the efficacy of the risk management function while impeding enterprise risk management; it increases operational inefficiency as consolidating data eats up bandwidth and requires extensive reconciliation and cleansing activities; and, finally, it decreases the profitability of a bank as cross-sell and up-sell opportunities are limited. Therefore, the collation and centralization of data driven by shared data across LOBs is an absolute imperative.


To mitigate the aforementioned consequences of siloed data, an effective data management strategy is the 3P method—prepare, process, and present. In the following sections, I will discuss the 3Ps of data management and reveal the value of a robust data management strategy.




Vemuri: “Banks must be prepared to handle humongous sets of data”

As we all know, banks store a tremendous amount of data across all lines of business—most of it hidden in disparate systems and unable to be easily consolidated. Hence, the first step in mobilizing data to better serve the needs of your bank is to prepare—or identify—the data offering utility or value. More times than not, the prepare step requires data validation or data quality tools. These tools allow users to enter in functional requirements to retrieve data sets relevant to their business needs.


The next challenge in the prepare step is to retrieve and consolidate data found across the bank. Fortunately, there are a bevy of data consolidation tools leveraging ETL (extract, transform, and load) technology to consolidate data across multiple disparate sources in real-time. ETL tools can extract data from sources as diverse as databases, mainframe files, Web services, information on the Internet and data on desktops within Microsoft Excel. In short, if there is data across your bank, ETL tools will find it.

"Banks will need to ensure proper investments are made in data management to squeeze all they can from their own cache of data"


The process step is the bread and butter of a robust data management strategy where analytical tools apply complex algorithms to large volumes of data producing a summarized and aggregated view of a data set. Running analytics on enterprise data sets is no small feat. The processing power required to complete analytical tasks is huge but recent advances in technology—especially distributed computing where the work can be distributed to multiple servers—have helpd. The distributed approach to computing and analytics was made famous by a Google methodology referred to as MapReduce. In the market today, there are various analytical software packages leveraging the MapReduce methodology including Apache Hadoop. Without such tools, high volume analytics would be next to impossible—especially if your bank would like to vet data in a timely manner.


Recently, advances in cloud computing security have enabled banks to leverage the cloud to conduct analytics outside of the on-premises model. However, despite strengthened data security in the cloud, including many cloud vendors obtaining SAS 70 compliance, I’m sure most banks are still a few years away from moving customer data to the cloud.



Once the data is prepared and processed, the present step is the last mile of the 3P methodology where visual representation of data is made available to business users. While such a step appears to be basic, proper representation of data drives the efficacy of decision making. If an analyst, teller, loan officer, or a CSR employee—to name a few—does not have a proper understanding of the data represented on their screen, they will be unable to serve a customer effectively. In a lot of ways, the proper visual representation of data is business critical. Fortunately, data visualization continues to improve with the advent of Rich Internet Application technologies like Flex, HTML5, and Silverlight


While my overview of the 3P model offered enough techno-babble for a month (at least), the bottom line is banks must be prepared to handle humongous sets of data to meet the expectations of their customers, effectively manage enterprise risk, and maintain profitability. Banks today certainly see the value in data. However, going forward, they will need to ensure proper investments are made in data management to squeeze all they can from their own cache of data. After that is done, it is up to a bank to determine what to do with its data. Besides, in the end, we all know it is worth the squeeze.